Skip Navigation


MBE Advance Access originally published online on October 5, 2007
Molecular Biology and Evolution 2007 24(12):2775-2786; doi:10.1093/molbev/msm212
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Material
Right arrow All Versions of this Article:
24/12/2775    most recent
msm212v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Fujita, M. K.
Right arrow Articles by Moritz, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fujita, M. K.
Right arrow Articles by Moritz, C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2007. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org

Research Articles

Multiple Origins and Rapid Evolution of Duplicated Mitochondrial Genes in Parthenogenetic Geckos (Heteronotia binoei; Squamata, Gekkonidae)

Matthew K. Fujita*,{dagger},{ddagger}, Jeffrey L. Boore*,{ddagger},§ and Craig Moritz*,{dagger}

* Department of Integrative Biology, University of California, Berkeley
{dagger} Museum of Vertebrate Zoology, University of California, Berkeley
{ddagger} Joint Genome Institute and Lawrence Berkeley National Laboratory, Walnut Creek, CA
§ Genome Project Solutions, Hercules, CA

E-mail: mkfujita{at}berkeley.edu.


   Abstract
TOP
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 Supplementary Material
 Acknowledgements
 References
 
Accumulating evidence for alternative gene orders demonstrates that vertebrate mitochondrial genomes are more evolutionarily dynamic than previously thought. Several lineages of parthenogenetic lizards contain large, tandem duplications that include rRNA, tRNA, and protein-coding genes, as well as the control region. Such duplications are hypothesized as intermediate stages in gene rearrangement, but the early stages of their evolution have not been previously studied. To better understand the evolutionary dynamics of duplicated segments of mitochondrial DNA, we sequenced 10 mitochondrial genomes from recently formed (~300,000 years ago) hybrid parthenogenetic geckos of the Heteronotia binoei complex and 1 from a sexual form. These genomes included some with an arrangement typical of vertebrates and others with tandem duplications varying in size from 5.7 to 9.4 kb, each with different gene contents and duplication endpoints. These results, together with phylogenetic analyses, indicate independent and frequent origins of the duplications. Small, direct repeats at the duplication endpoints imply slipped-strand error as a mechanism generating the duplications as opposed to a false initiation/termination of DNA replication mechanism that has been invoked to explain duplications in other lizard mitochondrial systems. Despite their recent origin, there is evidence for nonfunctionalization of genes due primarily to deletions, and the observed pattern of gene disruption supports the duplication–deletion model for rearrangement of mtDNA gene order. Conversely, the accumulation of mutations between these recent duplicates provides no evidence for gene conversion, as has been reported in some other systems. These results demonstrate that, despite their long-term stasis in gene content and arrangement in some lineages, vertebrate mitochondrial genomes can be evolutionary dynamic even at short timescales.

Key Words: mitochondrial genome • mtDNA • molecular evolution • Heteronotia • tandem duplication • parthenogenesis


   Introduction
TOP
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 Supplementary Material
 Acknowledgements
 References
 
Consistent content and streamlined organization are pervasive patterns of the animal mitochondrial genome that reflect constraints on its evolution. In addition to intracellular selection favoring efficient DNA replication (Rand 2001Go), a recent study has implicated faster mutation rates as an important factor in reducing the amount of noncoding DNA present in mitochondrial genomes (Lynch et al. 2006Go). Although these processes explain the consistent macroevolutionary pattern of reduced mitochondrial genome size, recent discoveries of alternative gene arrangements in birds (Mindell et al. 1998Go), reptiles (e.g., Kumazawa et al. 1996Go; Macey et al. 1997Go; Rest et al. 2003Go), and amphibians (e.g., Liu et al. 2005Go; Mueller and Boore 2005Go; San Mauro et al. 2006Go), among others, suggest that the conceived "static" mitochondrial genome is actually quite evolutionarily dynamic. Mounting evidence supports a "duplication–random loss" model of mitochondrial genome organization, in which the random deletion of duplicated genes can result in a gene order rearrangement (Moritz et al. 1987Go; Boore 2000Go; San Mauro et al. 2006Go). Yet, despite an increasing number of known alternative gene orders, it has been quite rare to observe mitochondrial genomes with the intermediate state of degenerating duplications (but see Macey et al. 1998Go; Mueller and Boore 2005Go; San Mauro et al. 2006Go).

We have previously documented the frequent occurrence of aberrant mitochondrial genome structures in asexual squamate lineages (geckos [Gekkonidae: Heteronotia binoei; Moritz 1991Go; Zevering et al. 1991Go; Hemidactylus garnotti; Moritz, unpublished data], whiptail lizards [Teiidae: Aspidoscelis; Moritz and Brown 1987Go], and rock lizards [Lacertidae: Darevskia; Moritz, unpublished data]). Although the gene content remains typical of all other animals, these mitochondrial genomes have a high incidence of tandem duplications of up to 10.4 kb, encompassing tRNA genes, rRNA genes, protein-coding genes, and the control region (Moritz and Brown 1987Go; Moritz 1991Go; Zevering et al. 1991Go). Stem-loop structures associated with the duplication boundaries in Aspidoscelis likely play an important role in generating these duplications, though the presence of direct sequence repeats can also generate tandem duplications by slipped-strand mispairing (Stanton et al. 1994Go). These features offer unique opportunities to investigate the tempo and mode of mitochondrial genome evolution.

Parthenogenetic geckos belonging to the H. binoei complex have mitochondrial genomes that range in size from a typical 17 kb to more than 27 kb as a result of tandem duplications that span tRNAs, rRNAs, protein-coding genes, and the control region (Moritz 1991Go; Zevering et al. 1991Go). There are 2 mitochondrial lineages of parthenogenetic Heteronotia that originated by reciprocal hybridization events between the 2 sexual chromosomal forms "CA6" and "SM6": the older 3N1 parthenogenetic lineage has a "CA6" maternal (mitochondrial) ancestor, whereas the more recent, independently derived 3N2 parthenogenetic lineage has an "SM6" maternal ancestor (fig. 1; Moritz 1993Go; Strasburg and Kearney 2005Go). Although tandem duplications within the mitochondrial genomes have arisen independently in both the 3N1 and 3N2 lineages, they have not been found in related sexual lineages despite extensive screening (Moritz 1991Go; Moritz and Heideman 1993Go). The low sequence diversity of nonduplicated mtDNA suggests that the parthenogenetic lineages formed within the last 300,000 years (Kearney et al. 2006Go), setting the maximum age of the mtDNA duplication events unique to these lineages. Mitochondrial genes in the 3N2 parthenogenetic forms evolve as expected when they exist in duplicate: 1 gene copy remains active and functional, whereas deletions and base-pair mutations render the other copy a pseudogene (Zevering et al. 1991Go). Unlike the 3N2 forms, which likely experienced a single mtDNA duplication event in their history, some 3N1 geckos have no duplications, whereas others exhibit a diversity of duplications, from 1.2 kb up to 10.4 kb (Moritz 1991Go). Restriction fragment analyses of the 3N1 genomes revealed some deletion events that presumably destroyed the functionality of the affected gene copies, resulting in pseudogenes residing within the mitochondrial genome (Moritz 1991Go). However, these restriction fragment length polymorphism (RFLP) analyses lacked the resolution necessary to fully investigate questions regarding the tempo and mode of mtDNA duplication evolution.


Figure 1
View larger version (8K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 1.— Origin of (A) 3N1 and (B) 3N2 parthenogenetic geckos of the Heteronotia binoei complex. The bold italics follow the maternal ancestry of the lineages.

 
This study describes complete sequences for 11 mitochondrial genomes, 1 of the CA6 type and 10 of the 3N1 types. Of the 10 3N1 mtDNAs, 6 harbor duplications ranging in size from 5.7 to 9.4 kb. We investigate the origins and evolution of the duplicated segments by 1) determining the structure of duplication endpoints and the junction point between duplication copies, 2) identifying mutations that have caused nonfunctionalization, 3) quantifying divergence between genomes and between duplicated regions within genomes, and 4) testing for single versus multiple origins of duplications in the 3N1 system on an inferred phylogeny.


   Materials and Methods
TOP
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 Supplementary Material
 Acknowledgements
 References
 
Sequencing and Annotating the Mitochondrial Genomes
We sequenced 11 mitochondrial genomes: 1 CA6 (the sexual maternal ancestor), 4 3N1 genomes without any duplications (samples 0A–0D), and 6 genomes with duplications of varying sizes (samples 8.8A, 8.8B, 7.2, 6.3, 9.4, and 5.7); we refer to the tandem duplications in each genome as copy 1 and copy 2 (table 1 and fig. 2). We refer to these genomes based on the size of their duplication as predicted by restriction digest analysis (Moritz 1991Go). We included 2 genomes with 8.8-kb duplications in order to compare mutation accumulation over short as well as longer timescales. The geographic distribution of our sampling spans the range of the 3N1 lineages in Western Australia and South Australia (table 1).


View this table:
[in this window]
[in a new window]

  		Table 1 Information and Mitochondrial Genome Attributes of the Samples Used in This Study

 

Figure 2
View larger version (48K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 2.— Linear representations of the mitochondrial genomes from this study. The genomes are shown arbitrarily to start at cytochrome oxidase I (cox1). Dark gray and light gray bars above the genomes represent the tandem duplications; numbers above the bars indicate the size of the duplicated copy. Genes outlined in bold are pseudogenes with arrows indicating nonfunctionalizing deletions (the numbers represent the location of the deletion within that gene) and lines indicating nonfunctionalizing base point mutations. The black bar above the typical genome designates the region affected by the tandem duplications, which also corresponds to the region of the genome left in a single-stranded state for the longest period of time during mtDNA replication (see text). Arrows underneath the typical genome show the directions of heavy and light-strand DNA synthesis. CR, control region; OL, origin of light-strand synthesis.

 
Using as template CsCl-purified mtDNA extractions previously used for RFLP studies (Moritz 1991Go), we amplified entire mitochondrial genomes in overlapping fragments several kilobases in length using long polymerase chain reaction (PCR). Supplementary figure 1 (Supplementary Material online) illustrates the amplification scheme used for each genome, supplementary table 1 (Supplementary Material online) lists the primers and PCR conditions used to amplify each fragment, and supplementary table 2 (Supplementary Material online) provides primer sequences. We purposely amplified each duplicate copy separately in order to avoid the coassembly of paralogous segments. We employed 2 long PCR protocols to amplify the large fragments. The first used the GeneAmp XL PCR kit (Applied Biosystems, Foster City, CA) following the manufacturer's protocol. The second used long and accurate Taq DNA polymerase (Takara Bio Inc., Shiga, Japan) following the manufacturer's protocol. Our cycling conditions were initial denaturation at 94 °C for 3 min; 37 cycles of 94 °C for 30 s, annealing temperature 30 s, and 68 °C extension (see supplementary table 1, Supplementary Material online for specific annealing temperatures and extension times for each fragment); final extension at 72 °C for 20 min; and final hold at 4 °C. We visualized each amplification on a 1% agarose gel stained with ethidium bromide to verify the size of the PCR product.

We used a shotgun sequencing approach to determine the sequence of the long PCR fragments. Long PCR products were first purified by ethanol precipitation and resuspended in 100 µl of 10 mM Tris (pH 8). We sheared the PCR products into 1–2 kb fragments using a Hydroshear (Genomic Solutions Inc., Ann Arbor, MI), repaired the ends using T4 DNA polymerase and DNA polymerase I, and ligated the resulting blunt-ended fragments into EcoRV-cut and dephosphorylated pMCL200 vector (Nakano et al. 1995Go) or a SmaI-cut and dephosphorylated pUC19 vector. We used the ligations to transform electrocompetent DH10β cells (Invitrogen Co., Carlsbad, CA) according to the manufacturer's protocol and then plated the transformation onto Luria broth + agar + isopropyl beta-D-1-thiogalactopyranoside (IPTG) + X-Gal plates with the appropriate antibiotic (ampicillin for the pUC19 vector, chloramphenicol for the pMCL200 vector). After a 16-h incubation at 37 °C, we picked approximately 24 white colonies per 1 kb of the original long PCR fragment (e.g., for an 8-kb long PCR product, we sequenced ca. 192 colonies). These colonies were processed at the DOE Joint Genome Institute (detailed protocols at http://www.jgi.doe.gov/sequencing/protocols/prots_production.html) through rolling circle amplification of the plasmids (TempliPhi, GE Healthcare, Piscataway, NJ), sequencing using BigDye v3.1 terminator chemistry (ABI), purification using "solid phase reversible immobilization," and analysis on an ABI3730xl (ABI) to produce a sequencing read from each end of each clone.

We assembled reads from individual libraries using Phrap (Gordon et al. 1998Go) and visualized and edited the resulting contigs by manual inspection in either Consed (Gordon et al. 1998Go) or Sequencher v.4.5 (GeneCodes, Ann Arbor, MI). Once we had all parts of the genome complete in individual contigs, we assembled the contigs to generate the mitochondrial sequence using Sequencher. By assembling the genomes in a step-like fashion, we avoided the coassembly of duplicated regions. There were several instances, however, where fragments did not overlap. In such cases, we sequenced the gaps directly from the PCR products, as depicted in supplementary fig. 1 and table 1 (Supplementary Material online). We used the consensus sequence for annotation and downstream analyses.

We generated an annotation of each genome using the web-based organelle genome annotation tool Dual Organellar Genome Annotator (DOGMA) (Wyman et al. 2004Go), refining the annotation, and inspecting potential misidentification of genetic elements using published mitochondrial genome sequences of 2 other geckos (Gekko gekko and Teratoscincus keyserlingii, GenBank accession numbers AY282753 [GenBank] and AY753545 [GenBank] , respectively).

Stem-loop structures and direct repeats occurring within the vicinity of the junction point or duplication endpoints may represent signals initiating potential mechanisms generating the tandem duplications, such as alternative initiation and/or termination of DNA replication. We used the hybrid-ss-sim program (default parameters for DNA) in the UNAFold package to predict secondary structures within approximately 25 bp of each side of each duplication endpoint (UNAFold web server: http://www.bioinfo.rpi.edu/applications/hybrid/; Markham and Zuker 2005Go). To evaluate the significance of these structures, we calculated a "P value" following the general procedure as described in Lavrov et al. (2000)Go. We wrote a Perl script to generate 1,000 sequence permutations of each 50-bp endpoint examined using the program Shuffle in the Wisconsin Package v.10.1 (Genetics Computer Group [GCG]; Perl script is available from M. Fujita upon request). For each permutation, we determined the {Delta}G values for the secondary structure predicted by hybrid-ss-sim. The P value is the frequency of secondary structures whose {Delta}G values were equal to or more exothermic than the structure predicted from the actual sequence. The presence of direct repeats at the duplication endpoints is consistent with a slipped-strand mispairing mechanism to generate the tandem duplications (Stanton et al. 1994Go). Because of their usefulness to quickly identify repeating elements (Mount 2004Go), we used dot plots (window = 5 bp) to compare 1) the first 100 bp of the 5' end of duplication copy 1 and the last 100 bp of the 3' end of duplication copy 2, 2) the 100-bp sequences immediately flanking the duplicated region, and 3) whole mitochondrial genomes (window = 10 bp).

Divergence and Phylogenetic Analysis
We used MUSCLE v3.6 (Edgar 2004Go) to align 1) each gene, including the control region, to identify gene-specific mutations; 2) the tandem duplications, including homologous regions spanning the duplications from genomes without duplications (nad2–nad4 inclusive of the control region), for measuring divergence and for phylogenetic analyses; and 3) whole genomes, excluding "copy 2" from those genomes with duplications, for phylogenetic analyses. For each alignment, we used PAUP v4.10b (Swofford 2003Go) to calculate maximum likelihood (ML)–corrected distance using models chosen for each data set by Modeltest v3.7 (Posada and Crandall 1998Go).

We determined phylogenetic relationships between paralogs and homologous segments from nonduplicated 3N1 genomes (with CA6 serving as the outgroup) to test whether there was a single tandem duplication event early in the history of 3N1 or whether there were several independent duplication events. Under the "single event" hypothesis, the same copy from different genomes should form clades (there should be a "copy 1" clade and a "copy 2" clade) to the exclusion of 3N1 genomes without duplications (fig. 3A). Under the "multiple events" hypothesis, copies from within genomes should group exclusively from other genomes, and there is no restriction as to the placement of genomes without duplications (fig. 3B). Concerted evolution between duplicate copies could generate patterns similar to the multiple events hypothesis, but our data indicate that this likely does not occur between Heteronotia mitochondrial duplications (see Discussion). We tested these 2 hypotheses using the Shimodaira–Hasegawa (SH) test (Shimodaira and Hasegawa 1999Go), which compares the fit of a model to the ML tree and to the alternative trees. To do this, we determined ML trees for the 3 scenarios (as-is stepwise addition and Tree Bisection-Reconnection [TBR] branch swapping) using the model of sequence evolution picked based on the hierarchical likelihood ratio test (hLRT) from Modeltest for the duplications alignment (Hasegawa-Kishino-Yano [HKY] + {Gamma} + I: Ti:Tv = 4.4888; shape parameter = 0.8346; proportion of invariant sites = 0.6946; frequency A = 0.3426, C = 0.3010, G = 0.1276; Hasegawa et al. 1985Go). These 3 trees included: the ML tree (fig. 3C), a tree constrained according to the single event hypothesis (fig. 3A), and a tree constrained to the multiple event hypothesis (fig. 3B). We used the SH test as implemented in PAUP v4.10b (Swofford 2003Go), with 1,000 bootstrap replicates and full optimization, to determine whether the ML tree differed significantly from the constrained trees. One potential pitfall of the SH test is that it is conservative (Goldman et al. 2000Go; Buckley 2002Go; Shimodaira 2002Go); therefore, we also performed the approximately unbiased (AU) test as implemented in the program CONSEL v0.1i (Shimodaira and Hasegawa 2001Go). To assign confidence to nodes in the ML tree, we performed a bootstrap analysis with 1,000 replicates. As a further test of the multiple events hypothesis, we performed an additional ML analysis with 1,000 bootstrap replicates using whole mitochondrial genomes (excluding duplicate copy 2 from the genomes with duplications, as well as ambiguously aligned, repeating segments from the control region; fig. 3D) using the Tamura-Nei [TrN] + {Gamma} model as chosen by the hLRT in Modeltest 3.7 (shape parameter = 0.0101; frequency A = 0.3243, frequency C = 0.3114, frequency G = 0.1301; rate matrix = (A–C = 1. A–G = 14.5824, A–T = 1, C–G = 1, C–T = 10.1757); Tamura and Nei [1993Go]). We also performed a 1,000-bootstrap maximum parsimony analysis (TBR branch swapping and as-is stepwise addition). We used PAUP v4.10b for all analyses (Swofford 2003Go).


Figure 3
View larger version (24K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 3.— (A) Phylogenetic relationships of duplication copies constrained under the single-duplication hypothesis, where one duplication event occurred before diversification. The origin of parthenogenesis and duplication events are mapped onto the phylogeny. (B) Phylogenetic relationships of duplication copies constrained under the multiple-duplications hypothesis, where several duplications occurred independently in different lineages. (C) ML tree of the phylogenetic relationships of the duplications with bootstrap support values (50% majority rule consensus of 1,000 bootstrap replicates). (D) ML tree of the phylogenetic relationships of mitochondrial genomes with bootstrap support values (50% majority rule consensus of 1,000 bootstrap replicates).

 
To quantify mutation accumulation, we used MacClade v4.08 to map unambiguous changes on the ML phylogeny of the duplications (Maddison and Maddison 2005Go). We used a binary system to code insertions and deletions; indels larger than 1 bp were considered a single indel. For superimposed (nested) indels, it was impossible to determine the character state of the smaller indel; in such cases, we coded the character state as missing data for the sample with the larger indel. This exercise provides an overall summary of the changes that have occurred within each of the duplications.

All data, including genome sequences, alignments, and tRNA secondary structures are available from M. Fujita. Genome sequences are also available on GenBank (accession numbers EF626807 [GenBank] –EF626817 [GenBank] ).


   Results
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 Supplementary Material
 Acknowledgements
 References
 
The mitochondrial genomes from Heteronotia have the typical vertebrate gene order (e.g., the "typical" genomes in fig. 2) and range in size from 16,988 bp (CA6) to 26,474 bp (9.4) (table 1). In genomes with duplications, each copy also follows the typical vertebrate gene order. The sizes of the duplications closely reflect estimates based on RFLP analysis (Moritz 1991Go). Sizes of duplicate copies within a given genome are similar except in 6.3, which has a 8023-bp copy 1 and a smaller 6467-bp copy 2 due mainly to a large, 1465-bp deletion spanning the 3'-end of nad5 and the 3'-end of nad6 (reverse complimented). Thus, the original duplication was >8 kb rather than 6.3 kb as inferred from RFLP analysis in Moritz (1991)Go. In 8.8A, the 2 duplicate copies are slightly different sizes due to a 297-bp deletion in nad6 copy 2. Otherwise, after the removal of ambiguously aligned portions of the control region (mostly short tandem repeats), the 2 copies present in each duplication are nearly identical in size.

The junction point—the point between the tandem duplications—differs among genomes except for 8.8A and 8.8B (fig. 4). Indeed, the gene combinations represented at the junction (junction genes) differ across genomes, though nad1 contributed to 4 junction points (5.7, 6.3, 8.8A, and 8.8B), and nad4 and nad5 each contributed to 2 junctions (8.8A, 8.8B and 6.3, 9.4, respectively); other junction genes include cob (5.7), nad2 (9.4), nad6 (7.2), and trnQ (7.2). For a given genome, the junction genes have the same or similar sequence where they meet at the junction point (fig. 4). For instance, the junction point in the 8.8 genomes is the sequence 5'-CCTGCACTT-3', which is present in both nad1 and nad4 at the point where they meet (fig. 4). Because of the tandem nature of these duplications, this shared sequence also occurs at each of the duplication endpoints and could potentially serve as the direct repeats implicated in the slipped-strand mispairing model of duplication. Such direct repeats (sometimes imperfect) are evident proximal to the junction in all duplications, except for 6.3. Our dot plot analysis did not reveal additional evidence of direct repeats within 100 bp of the duplication endpoints. A dot plot of the whole mitochondrial genomes did find a variety of matching repeats of similar length and percent similarity as those found at the duplication endpoints. However, the near-consistent presence of these repeats at the endpoints implies a connection to duplication origin. Because secondary structures may also play a role in the generation of the tandem duplications, we looked for potential stem-loop structures within the vicinity of the duplication endpoints. Though several of the predicted stem-loop structures were favorable ({Delta}G < 0), none of them were significant according to our randomization analysis (fig. 4).


Figure 4
View larger version (32K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 4.— Occurrence of secondary structures and direct repeats at the junction and endpoints of the tandem duplications. Illustrated for each genome is an alignment between the sequence at the junction (middle sequence) and the endpoint sequences (top and bottom sequences); asterisks represent aligned base pairs. Boxed sequences indicate perfect or near-perfect direct repeats and their matches at the junction. Shared underline patterns indicate regions that form potential stem structures as predicted by the program hybrid-ss-sim. The free energy values ({Delta}G) and the P value as calculated in our study (see Materials and Methods) are given for each endpoint structure.

 
There was very little divergence between the nonduplicated regions of the 3N1 mitochondrial genomes. This 9.8-kb region spans parts of nad2 to nad4, including atp6, atp8, cox1, cox2, cox3, and associated tRNA genes. ML distances ranged from 0.036% between 8.8A and 8.8B to 0.36% between 6.3 and 8.8B. Genes included within the duplications also showed little divergence due to substitutions. The ML-corrected distances between duplicated regions within genomes ranged from 0.035% (5.7) to 0.13% (8.8A) with a mean of 0.086%. (supplementary table 3, Supplementary Material online). Nonetheless, there were multiple mutations between paralogs within genomes, some of which caused nonfunctionalization (fig. 2, and see below).

To quantify patterns of sequence evolution following segmental duplication, we mapped unambiguous character changes on the ML phylogeny for the duplications (fig. 5). As expected from comparative analysis of nonduplicated mtDNA, transitions were more frequent than transversions (overall, 31 transitions vs. 7 transversion between pairwise comparisons of paralogs). In stark contrast to nonduplicated genomes (other than the control region), for which there was just 1 indel (43 bp in the rnl gene of 0D), indels were relatively common in the duplicated segments (n = 4 insertions, 14 deletions). In each case, only one of the pair of paralogous genes was affected by a specific indel.


Figure 5
View larger version (11K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 5.— The ML topology of the duplications. Mapped onto this phylogeny are unambiguous mutations (Ti, transition; Tv, transversion; i, insertion; and d, deletion).

 
Besides the gene truncations from involvement at the junction point, deletions were the most frequent mutations leading to nonfunctional genes (pseudogenes), as illustrated in figure 2. In 6.3, 283 bp missing from the 5' end of nad6 is part of a large 1465-bp deletion that spans both nad5 and nad6. A large deletion spans positions 182–501 in 8.8A nad6 copy 2. Both these deletions in the nad6 genes severely truncate and destroy the reading frame of the gene. The other nonfunctionalization mutations involve 1-bp frameshift deletions; these deletions are position 5 in 7.2 cob copy 2, positions 77 and 862 in 8.8A nad5 copy 1, position 78 in 8.8A nad6 copy 2, and position 77 in 8.8B nad5 copy 2. Two point mutations contribute to nonfunctionalization. An A -> G transition creates a premature stop codon at position 487 of nad1 copy 2 of the 9.4 genome, the more definitive instance of a base-pair mutation creating a pseudogene in our data set (figs. 2 and 5). A T -> C transition at the second position of the initiation codon in cob copy 1 in 9.4 causes an amino acid change (methionine to threonine) that may reduce or destroy cob functionality or push back translation initiation to the next available start codon (30 bp into the gene).

All tRNA genes can be folded into the typical stem-and-loop structures, and their gene order follows that typical for vertebrate mitochondrial genomes. Most of the nucleotide substitutions in the tRNA genes are between the CA6 genome and the 3N1 genomes, but there are also 3N1-specific substitutions as well as deletions. Within 3N1, substitutions occur in both stem regions (5 substitutions) and loop regions (1 substitution). It is difficult to assess whether any of these mutations, such as the 2-bp deletion in the first copy of trnH in 8.8A, caused pseudogenization. Because of its involvement in forming the junction point in 7.2, trnQ is missing most (32 bp) of its 3' end, a mutation that created the only instance of a definitive nonprotein-coding pseudogene.

The control region ranges in size from 1,554 bp (in duplication copy 2 from the 6.3 genome) to 2,100 bp (in duplication copy 2 of the 7.2 genome). Variation in size arises due to several regions with tandemly repeated sequences. Aside from 5.7, all other genomes with duplications had control regions of varying sizes (table 1). Most notable is the 370-bp size difference between 7.2 control region copy 1 (1730 bp) and control region copy 2 (2,100 bp) as a result of a large insertion. Prior to data analysis, we removed regions of ambiguous alignment caused by these tandem repeats. As with the rest of the genome, there is minimal substitutional divergence between 3N1 control regions both between copies within genomes and across genomes (supplementary table 3, Supplementary Material online). ML corrected pairwise distances between copies within genomes range from 0.065% (7.2) to 0.33% (8.8A) with a mean value of 0.14% and across genomes the corresponding values range from 0% to 0.65% (mean 0.30%).

We employed a ML phylogenetic approach to further evaluate whether the duplications arose via multiple events. (The maximum parsimony analysis produced nearly identical results but does not include the following tests of confidence.) The root is poorly resolved due to the low divergence among the 3N1 genomes (<0.4%) relative to the large divergence (4%) between 3N1 and CA6 genomes. The well-supported relationships occur near the tips, including the grouping of duplicate copies from the same genome (e.g., 9.4 copy 1 and copy 2 form a clade with 98% bootstrap proportions [BPs], 5.7 copy 1 and copy 2 with 94% BPs, and 6.3 copy 1 and copy 2 with 87% BPs) and a clade that includes the 8.8 genomes with a nonduplicated genome (0A) to the exclusion of all other 3N1 sequences (fig. 3C; –lnL = 16159.87035). The well-supported relationships in the phylogenetic tree based on whole mitochondrial genomes is largely consistent with the topology found using only the duplicated regions (fig. 3D), though the inclusion of additional data did not help to resolve the deeper nodes. The patterns exhibited by both trees are consistent with the multiple events hypothesis. Indeed, SH tests significantly rejected the hypothesis that the duplication ML tree and the single-event constrained tree (fig. 3A; –lnL = 16452.33444) equally fit the model of sequence evolution (P < 0.001), though could not reject the multiple events hypothesis (fig. 3B; –lnL = 16161.20154; P = 0.605). Similarly, the less conservative AU tests significantly rejected the equivalence of the ML tree and single-event tree (P < 0.001) but could not reject the multiple events hypothesis (P = 1).


   Discussion
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
Supplementary Material
 Acknowledgements
 References
 
General Features of 3N1 Mitochondrial Genomes
The 3N1 mitochondrial genomes exhibit a diversity of genome sizes (ranging from a typical 17 to >27 kb) as a result of tandem duplications. None of the genomes that now have different duplication sizes share the same junction between tandem copies, though the 6.3 genome evidently arose via a >8-kb duplication and subsequent copy-specific deletion. Most genes at the duplication endpoints are protein coding, except for trnQ in 7.2. Despite the independence of each duplication, they seem to include the same general area with the 5' end around the nad4/nad5/nad6 region across the control region to the 3' end within the vicinity of the origin of light-strand synthesis (OL; fig. 2).

The gene content of these duplications is similar to that of other duplications in vertebrate mitochondrial genomes. Mitochondrial duplications in unisexual whiptail lizards (Aspidoscelis) also include a similar region as in Heteronotia, but the endpoints of the Aspidoscelis duplications were frequently associated with tRNA genes and the 3' end of the control region (especially trnP and trnT; Moritz and Brown 1987Go; Stanton et al. 1994Go). The amphisbaenian Bipes biporus has a small duplication consisting of trnT and trnP situated between cob and the control region (Macey et al. 1998Go). Gene rearrangements in plethodontid salamanders (Stereochilus marginatus, Plethodon elongatus, Aneides flavipunctatus, and Aneides hardii) also generally involve the region between nad1 and nad5 (Mueller and Boore 2005Go). The generation of duplicated control regions present in several taxa (e.g., snakes [Kumazawa et al. 1996Go], agamid lizards [Amer and Kumazawa 2005Go], and birds [Mindell et al. 1998Go; Bensch and Härlid 2000Go; Eberhard et al. 2001Go]) likely involved duplications within this same general area. Gene rearrangement in the ranid frog Fejervarya limnocharis probably arose from a tandem duplication of the region spanning trnL(tag) to the control region that includes the genes nad5, nad6, and cob (Liu et al. 2005Go). In some of these cases, the control region may play an important role in generating the duplications, perhaps with errors in initiation or termination of heavy-strand DNA synthesis. However, situations in which the control regions reside within the duplication copies (such as in Heteronotia) require invoking an alternative duplication mechanism (see below). Thus, a constraint may exist that limits duplications to the region between the origin of light-strand synthesis (OL; fig. 2) and nad4 in vertebrates (excluding other genes such as cox1, cox2, cox3, atp6, and atp8), although there may not be a common mechanism driving these duplications in the different taxa (e.g., Mueller and Boore 2005Go).

Divergence versus Concerted Evolution of Paralogs
The low divergence observed between duplicate copies in the 3N1 mitochondrial genomes could result from the homogenizing effects of gene conversion. In fact, several reports have documented concerted evolution (via gene conversion) between duplicated control regions in diverse taxa, including snakes (Kumazawa et al. 1996Go), birds (Eberhard et al. 2001Go), Australasian agamid lizards (Amer and Kumazawa 2005Go), ticks (Black and Roehrdanz 1998Go), and sea cucumbers (Arndt and Smith 1998Go). A population-scale study of duplicated control regions in ostracods demonstrated that gene conversion in the mitochondrial genome can occur rapidly, perhaps with every replication cycle (Ogoh and Ohmiya 2007Go; but see Eberhard et al. 2001Go). However, in the 3N1 genomes, each duplicate copy has its own unique mutations that include nucleotide mutations, insertions, and deletions in both the control region and in other duplicated genes, several of which are very conspicuous and function crippling (>1,000-bp deletions). Control regions within genomes also exhibit copy-specific mutations, including a 370-bp insertion on 7.2 copy 2, and differ in their length by having different numbers of repeating motifs, patterns inconsistent with gene conversion on the level observed in other organisms. In addition, duplication copies in the 3N2 Heteronotia mitochondrial genomes exhibit even greater copy-specific divergences than reported here for 3N1 (Zevering et al. 1991Go), providing further evidence for the absence or inefficiency of gene conversion within this system. We cannot entirely dismiss gene conversion; it is possible that it occurs in the parthenogens but that the molecular mechanism is less efficient, perhaps as a result of disrupted coevolved gene complexes in gene conversion machinery resulting from the hybrid genome, or that the time since duplication is too recent compared with the timescale of gene conversion (Eberhard et al. 2001Go). However, given our data, it seems likely that the overall low divergence between duplicate copies in any given 3N1 mitochondrial genome, and in particular duplicated control regions, resulted from recent duplication origin rather than gene conversion.

Independence of the Duplications
The origin of the diversity seen in the duplication sizes originated in 1 of 2 ways: 1) a single, large duplication occurred in an ancestral 3N1 mitochondrial genome, followed by several deletion events or 2) multiple, independent duplication events occurred in several 3N1 lineages. Our data strongly support the latter hypothesis, as concluded by Moritz (1991)Go. The phylogenetic analysis and topology tests rejected the "single-duplication" hypothesis. Based on the topology of the ML tree, the 7.2 duplications appear the least consistent with an independent duplication, though the branch separating the paralogous sequences has low BPs, and restriction fragment analyses by Moritz (1991)Go actually supports an independent duplication event in 7.2.

Further evidence that the duplications occurred independently of each other comes from the distinct junctions between duplicate copies from the different 3N1 genomes. Under a single-duplication model, all the genomes should share a common internal junction. For instance, the similar junction between duplications in 3N2 mitochondrial genomes implies that a single duplication occurred in the history of that parthenogenetic lineage. However, in 3N1, genomes with different duplication sizes all had distinct junction points, which is inconsistent with the single-duplication hypothesis, but expected under the multiple-duplications hypothesis. Although it is possible that different junction points could arise by rapid and dramatic genomic deterioration at that specific point in each genome, this is unlikely given the recency of the duplications. These 2 lines of evidence—the phylogenetic analyses and the distinct junction points—form a strong argument for the independence of the duplication events among the 3N1 mitochondrial genomes. Given the evidence for independent origins of duplications examined here and a maximum age of the parthenogenetic lineages in which they arose (300 Kya; Strasburg and Kearney 2006), we estimate a minimum rate of duplication of 16.7 events per Myr. Moritz (1991)Go observed at least 5 additional duplication size classes that, assuming further independent origins, doubles the duplication rate to 33 events per Myr. Our minimum estimates of duplication rate are on the order of those estimated for the nuclear genome (Lynch and Connery 2000Go).

Duplication Mechanism
There are several mechanisms that could generate duplications in mitochondrial genomes. For instance, recombination at the shared endpoint sequences would excise a minicircle, which can then reintegrate randomly back into the genome. Although it is plausible that such a mechanism can generate tandem duplications, there is no reason to expect that an adjacent integration site would be favored, so it is much more likely to generate nontandem duplications as seen in some taxa (e.g., plethodontid salamanders [Mueller and Boore 2005Go]). Similarly, transposition would mostly generate nontandem duplications. The consistent tandem nature of the duplications seen in Heteronotia and other parthenogenetic lizards make intramolecular recombination and transposition an unlikely (but not altogether impossible) mechanism generating the duplications in the 3N1 genomes.

Other potential mechanisms involve errors of mtDNA replication (Boore 2000Go). Heavy-strand replication initiates within the control region and extends approximately two-thirds the length of the mitochondrial genome before hitting OL, where light-strand synthesis starts (for a review see Clayton 2003Go; fig. 2). As a consequence of this asymmetric replication, genes adjacent to the control region are left in a single-stranded state much longer than genes downstream from the OL (fig. 2). In addition, the region between the control region and the OL (including the rRNA genes) remains single stranded for a significant period as well. Interestingly, these regions that remain single stranded for the longest amount of time correspond to the regions involved in the duplication in Heteronotia as well as other systems (see above). The region that is single stranded for only a minimal amount of time, such as the cytochrome c oxidase genes and ATPase subunit genes, rarely occur in mitochondrial duplications. This pattern is consistent with 2 alternative mechanisms that exploit this feature of mtDNA replication: slipped-strand mispairing and alternative stem-loop light-strand replication initiation/termination (Stanton et al. 1994Go). The hybrid nuclear genomes of parthenogenetic Heteronotia and Aspidoscelis may increase the vulnerability of these systems to mtDNA duplications due to errors as a result of less efficient, disrupted coevolved cytonuclear replication machinery, similar to mismatched transcriptional cytonuclear complexes in the copepod Tigriopus californicus (Ellison and Burton 2006Go).

Stem-loop structures may play important roles in the mechanisms generating the tandem duplications by serving as alternative initiation signals of light-strand synthesis (Macey et al. 1997Go). Stanton et al. (1994)Go found a strong association of stem-loop structures located within 5 bp of the inferred duplication endpoints in Aspidoscelis mitochondrial duplications. These structures have stems at least 4 bp in length and were always thermodynamically stable ({Delta}G < 0). In contrast, for Heteronotia, we found no evidence of stable stem-loop structures that could serve as alternative initiation signals for light-strand synthesis in the endpoint genes.

Within a given 3N1 mitochondrial genome, the duplication endpoints mostly have the same or share very similar sequences (fig. 4). These shared endpoints may act as direct repeats in a slipped-strand mispairing mechanism in generating the duplications during DNA replication. This could occur, for example, as the elongating light-strand dissociates near nad1 and reanneals to a repeat that has already replicated, such as in nad4, producing a tandem repeat approximately 8.8 kb in length that includes the control region. In contrast, most of the Aspidoscelis mitochondrial genomes surveyed by Stanton et al. (1994)Go did not share endpoint sequences (except for Aspidoscelis opatae and Aspidoscelis exsanguis, whose mitochondrial genomes are likely derived from a common ancestor). Although the exact mechanism generating the duplications is not known for certain in either Aspidoscelis or Heteronotia, it appears likely that they originated in different ways, though consistently with the area of maximum heavy-strand displacement.

Nonfunctionalization and the Duplication–Random Loss Model of Mitochondrial Genome Rearrangement
In our sampling, as many as 17 pseudogenes have formed in the mitochondrial genomes of parthenogenetic 3N1 Heteronotia within the last 300,000 years. This number will likely rise with the identification of additional nonfunctionalizing mutations in other lineages (Moritz 1991Go). Given the independence and recency of the duplications and the large number of pseudogenes present in the 3N1 mitochondrial genomes, it is clear that nonfunctionalization of duplicated genes occurs rapidly. This proceeds mostly by deletion events, either as a consequence of the duplication process where genes at the junction become truncated or by deletions that have occurred after the duplication event in other genes. All but one of the deletions observed in the duplicated regions occurred after the duplication event (fig. 5), perhaps reflecting relaxed functional constraint of redundant genes and a deletion bias (Ophir and Graur 1997Go). Nonfunctionalization due to nucleotide substitutions was rare, and we observed only one clear instance of a substitution creating a pseudogene by forming a premature stop codon (nad2 copy 2 from 9.4); another substitution in 9.2 cob copy 1 delays translation initiation by 10 codons. In this regard, deletions have more influence in disrupting function in duplicated gene copies. It is also evident that deletions can rapidly deteriorate a gene, even when nucleotide divergence between duplicate copies is low. For example, for nad5 copy 2 in 6.3, only 163 bp remain after being truncated at both ends of the gene, yet the remaining sequence is identical to its paralog.

Mitochondrial genomes from CA6 and the nonduplicated 3N1 samples have a gene order consistent with the typical gene order from animals (Boore 1999Go). However, deviations from the typical gene order do exist, and an increasing number of studies demonstrate that such deviations are not uncommon (e.g., Boore 1999Go; Mueller and Boore 2005Go). The most widely accepted mechanism for gene order rearrangement involves duplicating a segment of the mitochondrial genome followed by random loss of genes from the duplicate copies (Boore 1999Go, 2000Go). In Heteronotia, we have shown that duplications and subsequent nonfunctionalization and gene deterioration can occur frequently and rapidly. In fact, we can predict eventual gene order rearrangement assuming that the duplications deteriorate until the genome is the typical 17-kb size. For instance, one potential gene order rearrangement between 8.8A and 8.8B involves nad5 and nad6. Initially after the tandem duplications, the gene order in both 8.8 genomes is nad5 (copy 1)–nad6 (copy 1)–nad5 (copy 2)–nad6 (copy 2) (fig. 2). In 8.8A, the first copy of nad5 and the second copy of nad6 experienced nonfunctionalization deletions (fig. 2). Eventual elimination of these genes would result in a gene order of nad6–nad5, whereas similar logic will show that 8.8B will retain the typical nad5–nad6 order. This example demonstrates that gene order rearrangement is possible even between very closely related genomes.

Evolutionary Implications
If mitochondrial duplications occur as frequently as suggested in this study, then the lack of these mutations in most vertebrates suggests they are highly deleterious and purged from the population quickly (but see Mueller and Boore [2005]Go). Parthenogenetic systems are unique in having a high incidence of duplication mutations that also may persist for thousands of years (e.g., 3N2 duplications; Zevering et al. 1991Go). This phenomenon may occur for several reasons. First, the hybrid nature of the nuclear genome may form suboptimal cytonuclear complexes necessary for proper mtDNA replication, causing an increased frequency of duplication mutations; that is, parthenogenetic systems are equally intolerant of mitochondrial duplications compared with sexual organisms, but duplication events simply occur more frequently in hybrid unisexuals (Zevering et al. 1991Go). However, medical research in humans has clearly demonstrated the existence of several such variants in human somatic cells, implying that even within a diploid, sexual, nonhybrid individual, a large variety of rearrangements occur with a surprisingly high frequency (though discovery of these mutations is biased by their medical importance; e.g., Howell and Smejkal 2000Go; Wallace 2005Go). Second, for parthenogenetic lineages, mtDNA has double the effective population size compared with an equivalently populous sexual population, and thus harbors more genetic diversity, including duplication polymorphisms. However, natural selection is stronger in larger populations and should be more effective at eliminating the defective genome (Hartl and Clark 2007).

We propose a third hypothesis that requires further investigation. If males present a selective constraint to keep mitochondrial genomes streamlined (perhaps to increase transcription efficiency for the energy-demanding cells), then that constraint disappears in parthenogenetic populations, allowing typically deleterious mutations to exist in a stable manner. Because the maternal inheritance of mtDNA limits male influence of mitochondrial evolution (Frank and Hurst 1996Go), this hypothesis applies only if the male-biased deleterious mutations also influence mitochondrial fitness. For instance, mitochondrial mutations that cause females to produce "bad sons," or even "bad sperm" if there is cryptic sexual selection, reduce female fitness and, presumably by association, mitochondrial fitness. Several studies have already implicated mtDNA quality with sperm function (e.g., Ruiz-Pesini et al. 1998Go; Spiropoulos et al. 2002Go; but see Pereira et al. 2007Go), and one intriguing study discovered several abnormal phenotypes in experimental "mtDNA-mutator" mice, one of which was a profound reduction in fertility of afflicted male mice when crossed with wild-type females (Trifunovic et al. 2004Go; Zeh and Zeh 2005Go). Thus, it seems feasible that aberrant mitochondrial genomes can affect female fitness through mitochondrial genome quality in males, though there needs to be theoretical evaluation of this hypothesis.

Several hundred sequenced animal mitochondrial genomes attest to their structural efficiency and consistent gene content. Recent theory argues that increased mutation pressure relative to larger, more complex genomes prevents animal mitochondrial genomes from having additional features such as introns and other noncoding segments of DNA with regulatory functions (Lynch et al. 2006Go). Indeed, the increasing number of observations of rearranged gene orders continues to build evidence for the duplication–random loss model, in which a transient phase of rapid degeneration of potentially deleterious extraneous sequence quickly reestablishes the characteristic streamlined feature of mitochondrial genomes (Moritz et al. 1987Go; Boore 2000Go; San Mauro et al. 2006Go). Here, using the unique features of Heteronotia mitochondrial genomes, we have shown that this process of nonfunctionalization and rapid deterioration of duplicated genes can occur frequently among closely related lineages.


   Supplementary Material
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Supplementary Material
Acknowledgements
 References
 
Supplementary figure 1 and tables 1–3 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).


   Acknowledgements
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Supplementary Material
 Acknowledgements
References
 
We thank members of the Moritz and McGuire labs for valuable discussion and comments on earlier drafts of this manuscript. A. Leache provided useful advice on the phylogenetic analyses. J.R. Macey, J. Fong, and W.K. Savage provided laboratory support and discussion early in the project. The South Australian Museum provided the tissue samples. We acknowledge the National Institute of Genetics (Japan) for providing the pMCL200 vector. N. Takezaki, D. San Mauro, and 2 anonymous reviewers provided important comments that improved the quality of this manuscript. This work was supported by the Department of Integrative Biology at University of California Berkeley, the Museum of Vertebrate Zoology, and an National Science Foundation Graduate Research Fellowship. This work was performed partly under the auspices of the US Department of Energy's Office of Science, Biological and Environmental Research Program, and by the University of California, Lawrence Berkeley National Laboratory under Contract No. DE-AC02-05CH11231.


   Footnotes
 
Naoko Takezaki, Associate Editor


   References
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Supplementary Material
 Acknowledgements
 References
 

    Amer SAM, Kumazawa Y. Mitochondrial genome of Pogona vitticepes (Reptilia; Agamidae): control region duplication and the origin of Australasian agamids. Gene (2005) 346:249–256.[CrossRef][Web of Science][Medline]

    Arndt A, Smith MJ. Mitochondrial gene rearrangement in the sea cucumber genus Cucumaria. Mol Biol Evol (1998) 15:1009–1016.[Abstract]

    Bensch S, Härlid A. Mitochondrial genomic rearrangements in songbirds. Mol Biol Evol (2000) 17:107–113.[Abstract/Free Full Text]

    Black WC, Roehrdanz RL. Mitochondrial gene order is not conserved in arthropods: prostriate and metastriate tick mitochondrial genomes. Mol Biol Evol (1998) 15:1772–1785.[Abstract]

    Boore JL. Animal mitochondrial genomes. Nucleic Acids Res (1999) 27:1767–1780.[Abstract/Free Full Text]

    Boore JL. The duplication/random loss model for gene rearrangement exemplified by mitochondrial genomes of deuterostome animals. In: Comparative genomics, computational biology series—Sankoff D, Nadeau J, eds. (2000) Vol. 1. Dordrecht (The Netherlands): Kluwer Academic Publishers. 133–147.

    Buckley TR. Model misspecification and probabilistic tests of topology: evidence from empirical data sets. Syst Biol (2002) 51:509–523.[CrossRef][Web of Science][Medline]

    Clayton DA. Mitochondrial DNA replication: what we know. IUBMB Life (2003) 55:213–217.[Web of Science][Medline]

    Eberhard JR, Wright TF, Bermingham E. Duplication and concerted evolution of the mitochondrial control region in the parrot genus Amazona. Mol Biol Evol (2001) 18:1330–1342.[Abstract/Free Full Text]

    Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res (2004) 32:1792–1797.[Abstract/Free Full Text]

    Ellison CK, Burton RS. Disruption of mitochondrial function in interpopulation hybrids of Tigriopus californicus. Evolution (2006) 60:1382–1391.[CrossRef][Web of Science][Medline]

    Frank S, Hurst LD. Mitochondria and male disease. Nature (1996) 383:224.[CrossRef][Medline]

    Goldman N, Anderson JP, Rodrigo AG. Likelihood-based tests of topologies in phylogenetics. Syst Biol (2000) 49:652–670.[Abstract/Free Full Text]

    Gordon D, Abajian C, Green P. Consed: a graphical tool for sequence finishing. Genome Res (1998) 8:195–202.[Abstract/Free Full Text]

    Hartl DL, Clark AG. Principles of population genetics (2007) Sunderland (MA): Sinauer Associates.

    Hasegawa M, Kishino H, Yano Y. Dating of the human-ape splitting by a molecular clock of mitchondrial DNA. J Mol Evol (1985) 22:160–174.[CrossRef][Web of Science][Medline]

    Howell N, Smejkal CB. Persistant heteroplasmy of a mutation in the human mtDNS control region: hypermutation as an apparent consequence of simple-repeat expansion/contraction. Am J Hum Genet (2000) 66:1589–1598.[CrossRef][Web of Science][Medline]

    Kearney M, Blacket MJ, Strasburg JL, Moritz C. Waves of parthenogenesis in the desert: evidence for the parallel loss of sex in a grasshopper and a gecko from Australia. Mol Ecol (2006) 15:1743–1748.[CrossRef][Medline]

    Kumazawa Y, Ota H, Nishida M, Ozawa T. Gene rearrangements in snake mitochondrial genomes: highly concerted evolution of control-region-like sequences duplicated and inserted into a tRNA gene cluster. Mol Biol Evol (1996) 13:1242–1254.[Abstract]

    Lavrov DV, Boore JL, Brown WM. The complete mitochondrial DNA sequence of the horseshoe crab Limulus polyphemus. Mol Biol Evol (2000) 17:813–824.[Abstract/Free Full Text]

    Liu Z-Q, Wang Y-Q, Su B. The mitochondrial genome organization of the rice frog, Fejervarya limnocharis (Amphibia: Anura): a new gene order in the vertebrate mtDNA. Gene (2005) 346:145–151.[CrossRef][Web of Science][Medline]

    Lynch M, Connery JS. The evolutionary fate and consequences of duplicate genes. Science (2000) 290:1151–1155.[Abstract/Free Full Text]

    Lynch M, Koskella B, Schaack S. Mutation pressure and the evolution of organelle genomic architecture. Science (2006) 311:1727–1730.[Abstract/Free Full Text]

    Macey JR, Larson A, Ananjeva NB, Fang Z, Papenfuss TJ. Two novel gene orders and the role of light-strand replication in rearrangement of the vertebrate mitochondrial genome. Mol Biol Evol (1997) 14:91–104.[Abstract]

    Macey JR, Schulte JA 2nd, Larson A, Papenfuss TJ. Tandem duplication via light-strand synthesis may provide a precursor for mitochondrial genomic rearrangement. Mol Biol Evol (1998) 15:71–75.[Abstract]

    Maddison DR, Maddison WP. MacClade Version 4.08 (2005) Sunderland (MA): Sinauer Associates.

    Markham NR, Zuker M. DINAMelt web server for nucleic acid melting prediction. Nucleic Acids Res (2005) 33:W5577–W5581.

    Mindell DP, Sorenson MD, Dimcheff DE. Multiple independent origins of mitochondrial gene order in birds. Proc Natl Acad Sci USA (1998) 98:10693–10697.

    Moritz C. Evolutionary dynamics of mitochondrial DNA duplications in parthenogenetic geckos, Heteronotia binoei. Genetics (1991) 129:221–230.[Abstract]

    Moritz C. The origin and evolution of parthenogenesis in the Heteronotia binoei complex: synthesis. Genetica (1993) 90:269–280.[CrossRef][Web of Science]

    Moritz C, Brown WM. Tandem duplications in animal mitochondrial DNAs: variation in incidence and gene content among lizards. Proc Natl Acad Sci USA (1987) 84:7183–7187.[Abstract/Free Full Text]

    Moritz C, Dowling TE, Brown WM. Evolution of animal mitochondrial DNA: relevance for population biology and systematics. Annu Rev Ecol Syst (1987) 18:269–292.[CrossRef][Web of Science]

    Moritz C, Heideman A. The origin and evolution of parthenogenesis in Heteronotia binoei (Gekkonidae): reciprocal origins and diverse mitochondrial DA in western populations. Syst Biol (1993) 42:293–306.[Abstract/Free Full Text]

    Mount DW. Bioinformatics: sequence and genome analysis (2004) 2nd ed. Sunderland (MA): Sinauer Associates.

    Mueller RL, Boore JL. Molecular mechanism of extensive mitochondrial gene rearrangement in plethodontid salamanders. Mol Biol Evol (2005) 22:2104–2112.[Abstract/Free Full Text]

    Nakano Y, Yoshida Y, Yamashita Y, Koga T. Construction of a series of pACYC-derived plasmid vectors. Gene (1995) 162:157–158.[CrossRef][Web of Science][Medline]

    Ogoh K, Ohmiya Y. Concerted evolution of duplicated control regions within an ostracod mitochondrial genome. Mol Biol Evol (2007) 24:74–78.[Abstract/Free Full Text]

    Ophir R, Graur D. Patterns and rates of indel evolution in processed pseudogenes from humans and murids. Gene (1997) 205:191–202.[CrossRef][Web of Science][Medline]

    Pereira L, Gonçalves J, Franco-Duarte R, Silva J, Rocha T, Arnold C, Richards M, Macaulay V. No evidence for an mtDNA role in sperm motility: data from complete sequencing of asthenozoospermic males. Mol Biol Evol (2007) 24:868–874.[Abstract/Free Full Text]

    Posada D, Crandall KA. Modeltest: testing the model of DNA substitution. Bioinformatics (1998) 14:817–818.[Abstract/Free Full Text]

    Rand DM. The units of selection on mitochondrial DNA. Annu Rev Ecol Syst (2001) 32:415–448.[CrossRef][Web of Science]

    Rest JS, Ast JC, Austin CC, Waddell PJ, Tibbetts EA, Hay JM, Mindell DP. Molecular systematics of primary reptilian lineages and the tuatara mitochondrial genome. Mol Phylogenet Evol (2003) 29:289–296.[CrossRef][Web of Science][Medline]

    Ruiz-Pesini E, Diez C, Lapena AC, Preze-Martos A, Montoya J, Alvarez E, Arenas J, Lopez-Perez MJ. Correlation of sperm motility with mitochondrial enzymatic activities. Clin Chem (1998) 44:1616–1620.[Abstract/Free Full Text]

    San Mauro D, Gower DJ, Zardoya R, Wilkinson M. A hotspot of gene order rearrangement by tandem duplication and random loss in the vertebrate mitochondrial genome. Mol Biol Evol (2006) 23:227–234.[Abstract/Free Full Text]

    Shimodaira H. An approximately unbiased test of phylogenetic tree selection. Syst Biol (2002) 51:492–508.[CrossRef][Web of Science][Medline]

    Shimodaira H, Hasegawa M. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Mol Biol Evol (1999) 16:1114–1116.[Web of Science]

    Shimodaira H, Hasegawa M. CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics (2001) 17:1246–1247.[Abstract/Free Full Text]

    Spiropoulos J, Turnbull DM, Chinnery PF. Can mitochondrial DNA mutations cause sperm dysfunction? Mol Hum Reprod (2002) 8:719–721.[Abstract/Free Full Text]

    Stanton DJ, Daehler LL, Moritz CC, Brown WM. Sequences with the potential to form stem-and-loop structures are associated with coding-region duplications in animal mitochondrial DNA. Genetics (1994) 137:233–241.[Abstract]

    Strasburg JL, Kearney M. Phylogeography of sexual Heteronotia binoei (Gekkonidae) in the Australia arid zone: climatic cycling and repetitive hybridization. Mol Ecol (2005) 14:2755–2772.[CrossRef][Medline]

    Swofford DL. PAUP*: phylogenetic analysis using parsimony (*and other methods). Version 4 (2003) Sunderland (MA): Sinauer Associates.

    Tamura K, Nei M. Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol Biol Evol (1993) 10:512–526.[Abstract]

    Trifunovic A, Wredenberg A, Falkenberg M. 13 co-authors. Premature ageing in mice expressing defective mitochondrial DNA polymerase. Nature (2004) 429:417–423.[CrossRef][Medline]

    Wallace DC. A mitochondrial paradigm of metabolic and degenerative diseases, aging, and cancer: a dawn for evolutionary medicine. Annu Rev Genet (2005) 39:359–407.[CrossRef][Web of Science][Medline]

    Wyman SK, Jansen RK, Boore JL. Automatic annotation of organellar genomes with DOGMA. Bioinformatics (2004) 20:3252–3255.[Abstract/Free Full Text]

    Zeh JA, Zeh DW. Maternal inheritance, sexual conflict and the maladapted male. Trends Genet (2005) 21:281–286.[CrossRef][Web of Science][Medline]

    Zevering CE, Moritz C, Heideman A, Sturm RA. Parallel origins of duplications and the formation of pseudogenes in mitochondrial DNA from parthenogenetic lizards (Heteronotia binoei; Gekkonidae). J Mol Evol (1991) 33:431–441.[CrossRef][Web of Science][Medline]

Accepted for publication September 26, 2007.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Mol Biol EvolHome page
M. Dowton, S. L. Cameron, J. I. Dowavic, A. D. Austin, and M. F. Whiting
Characterization of 67 Mitochondrial tRNA Gene Rearrangements in the Hymenoptera Suggests That Mitochondrial tRNA Gene Position Is Selectively Neutral
Mol. Biol. Evol., July 1, 2009; 26(7): 1607 - 1617.
[Abstract] [Full Text] [PDF]

Home page
 Mol Biol EvolHome page
M. M. Fonseca, D. Posada, and D. J. Harris
Inverted Replication of Vertebrate Mitochondria
Mol. Biol. Evol., May 1, 2008; 25(5): 805 - 808.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Material
Right arrow All Versions of this Article:
24/12/2775    most recent
msm212v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Fujita, M. K.
Right arrow Articles by Moritz, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Fujita, M. K.
Right arrow Articles by Moritz, C.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?